Saylor Started MicroStrategy's Fire Himself

“MicroStrategy has become bitcoin’s Federal Reserve — this drop to $59,000 was a test of the Michael Saylor put.” — Zheng Di

Guest: Liu Feng｜host of the podcast Web3 101　·　Zheng Di｜investor, long-time reader of MicroStrategy’s 8-Ks

This episode was recorded on Sunday afternoon, June 21, 2026. Bitcoin has fallen from $126,000 at the end of last year to $58,900, close to halving; two days ago STRC, MicroStrategy’s hottest financing instrument this year, fell from its $100 par to a low of $82.50. A few days before that, MicroStrategy did something it had never done in its history: sold coins — though only 32 of them.

This is a crossover between Mossfire and Web3 101, hosted by Liu Feng, with two investors who have spent years reading MicroStrategy’s 8-K filings taking the questions: Raymond, net short bitcoin since the start of the year, and old friend Zheng Di, who knows STRC’s liquidation seniority and coupon trajectory cold. Zheng Di doesn’t think STRC is a Ponzi — he actually thinks it’s a beautiful piece of financial innovation — but he says outright that Saylor, whom he has always liked, made two errors in a row, starting with a $1.5 billion convertible buyback that appeared to net nearly $100 million. Raymond goes further: in a bear market, Saylor himself is the rat that knocked over the oil jar. Monday’s pre-market 8-K will give the first answer.

What follows is the full conversation, edited and condensed.

1. Any asset without growth gets pinned down by rates

Liu Feng: Every time we discuss MicroStrategy, we’re discussing whether it’s a great innovation or a Ponzi scheme. MicroStrategy’s stock fell last session and STRC has been very volatile. Raymond, open us up: why bring this topic out again now?

Raymond: MicroStrategy plays different roles at different historical stages. In a bull market it steps on its own feet all the way up, a perpetual motion machine, our strongest teammate; now bitcoin is close to halving and it’s stepping on its own feet all the way down, and has even started selling coins. This company has extremely strong reflexivity. Discussing MicroStrategy is really discussing macro liquidity, US capital markets’ tolerance for innovation, the power of retail, and bitcoin’s price itself. It’s a mirror.

Liu Feng: One line of Raymond’s struck me: every past bear-market bottom had a landmark event that shocked everyone, like FTX last cycle; this bear market seems confirmed, but the explosive event hasn’t appeared. Could MicroStrategy be that trigger?

Zheng Di: In a traditional framework, this cycle is still clearly rate suppression. I suspect the US economy is being underestimated — it may simply be too strong, with carbon-based and silicon-based diverging sharply. There were market rumors last year that ninety percent of the GDP increment came from AI infrastructure, and not long ago the Peterson Institute published a piece saying existing GDP accounting severely underestimates the contribution of tokens and the AI industry. So current inflation pressure and rate-hike expectations, beyond oil prices and geopolitical noise, are largely AI infrastructure propping the economy up too strongly — rate-cut expectations are gone and hikes are brewing instead. In that environment any asset without strong growth expectations is dull; bitcoin and gold both. Gold has its own story of course: after the 2022 Russia-Ukraine war it gradually escaped the Fed’s rate cage, and the recent strait closure and Middle Eastern countries being forced to sell gold to fund the balance of payments also matter. In short, apart from the AI supply chain — which frankly is now only memory — everything is falling, and the one-nine market, where a tenth rises and nine-tenths fall, is very pronounced. Within AI there’s differentiation too: the Magnificent Seven aren’t rising much, power and neoclouds are okay on tight compute, GPUs are rising in price, optics is being pressed by rumors of Chinese capacity expansion, and the only genuinely firm thing is memory. So-called AI stealing everyone’s thunder is a surface reading; the deeper reason is that memory’s growth expectations are strong enough not to fear the rate cycle. Anything whose growth isn’t strong enough gets pinned down. As for black swans, I have no answer. MicroStrategy’s recent moves genuinely were errors, but there’s no evidence at present that it blows up.

Raymond: I basically agree with Mr Zheng. Let me disclose the position first: since the start of this year I’ve been essentially net short bitcoin, and what I’m thinking about is when this short should be covered and in what way the cycle bottoms. My subjective feeling: the depth of the remaining fall isn’t necessarily great, but the correction may still take about six months. When liquidity is short, if A rises then B can’t, and bitcoin as the weaker asset becomes what you see today. As for blowing up — there may not literally be a “blow-up,” but its capital-markets operations could affect the coin price, and the coin price could then indirectly maim others. Last cycle, before FTX there was Three Arrows, and before Three Arrows there was DCG — did DCG actually blow up? Not really, but it pulled the chain behind it. MicroStrategy is exactly that kind of indicator worth watching.

2. The error was that $1.5 billion buyback: picking up sesame, dropping the watermelon

Liu Feng: Mr Zheng, concretely, where did MicroStrategy go wrong recently?

Zheng Di: The main misstep isn’t selling coins to pay interest — they’d been planning that a long time. Look, BitMine recently issued its own preferred too, and Tom Lee had a slight dig at MicroStrategy: bitcoin has no carry, so it isn’t suited to preferred issuance; Ethereum has staking yield and can cover the dividend. He’s right about one thing: turning an asset with no yield into a perpetual paying cash flow requires superb operating technique. MicroStrategy had been planning to sell coins since the start of the year, the signal at the May earnings call was already clear, and the odds on “sells coins this year” on Polymarket and Kalshi were both over 60%. None of that is a problem. But he made two misjudgements. First: if you’re going to sell coins to pay interest and merely run a demonstration project, then the dollar cash reserve should be maintained at the guided 12 to 24 months — before redeeming that $1.5 billion of converts, it was genuinely above 24 months, as high as 30 at one point, and nobody questioned the ability to pay. Then he got greedy about the seven or eight point discount on the converts, made over $90 million on the buyback, and the cash reserve was left at eight or nine hundred million, 7.7 months of coverage. Now every piece of FUD says: your cash is only nine hundred million or a billion, only enough for seven-odd months; the coupon is 11.5% and the market already demands 13%, so the moment you raise it, coverage falls to 7.3 or 7.4 months — a vicious cycle. Selling 32 coins, invoking a “bitcoin-per-share neutral” principle, and having a low cash reserve: those three cannot coexist. The error starts with that $1.5 billion buyback: picking up sesame, dropping the watermelon. That’s a strategic-level error.

Liu Feng: The core is that the buyback sharply reduced the cash reserve, causing market panic.

Zheng Di: Right. The second misjudgement is timing. In the window where he sold coins, liquidity expectations for global risk assets were already poor, the one-nine market had gone to an extreme, and it was right next to SpaceX’s listing drawing blood. With Saylor’s brain and his Wall Street connections, everything would have told him this wasn’t the window for a financial experiment. In May, when he talked about possibly selling small amounts and about the neutrality principle, the market didn’t believe it and assumed he’d pick his moment, so it didn’t fall — including me — and then our glasses shattered on the floor. By May’s arithmetic, selling stock to pay interest only made sense with mNAV above 1.22x; by the time he actually sold coins the threshold had risen to 1.29–1.3x, and he was only at 1.13x. So for the foreseeable future he’s permanently below his own threshold, and the market naturally expects him to be forced to sell coins every week, and immediately goes to test the floor. Where is that floor? MicroStrategy has become bitcoin’s Federal Reserve: every time the stock market falls, it’s to force out the Fed put; this drop to $59,000 was a test of the Saylor put. For it to stop falling you need to see a de facto soft pivot. Sure enough, a week later he announced selling stock, bought 1,800 coins and put $100 million of cash into the reserve. And here’s another small error: there was no need at all to buy 1,800; buying back the 32 he sold, or buying 100 as a statement, would have been enough, and the money should have gone to the reserve. But the coin price did stabilize, because the market finally believed he doesn’t treat bitcoin as an ATM. The cost was STRC collapsing: in this system, the share price, the coin price and STRC are three balls that must stay in the air continuously. You have money and don’t provision the dividend, and instead break precedent to sell stock and buy coins — STRC holders discover their priority has been ranked behind hoarding coins, and a sharp fall is entirely normal. Chained ships, three things bound together, the difficulty went up, but it isn’t unplayable. The sharp fall in the cash reserve is the fuse for burning the chained fleet — Saylor lit a fire under himself.

Raymond: Strongly agree. MicroStrategy is a company that lives extremely heavily on market confidence and retail confidence. Previously it was “I sell when I want to and don’t when I don’t,” with the initiative in hand; now it’s become “I may have no choice but to sell” — for a company that lives on confidence, the backbone has been pulled out. One point where I differ slightly: bitcoin’s recent stabilization looks more like a short rebound after falling too fast, with volume not reflecting real buying; the worry is still there. Windows for selling coins in a bear market are extremely few and the error tolerance is extremely low: in 2026 bitcoin was near $80,000 for only a handful of days — when he announced he’d sell coins the price was still at the relative high of around $80,000, and I thought it would crash and it didn’t, and I marveled at how strong bitcoin was; then over the following month it slid down. This isn’t over. A hundred-odd million a month of interest converts to only a thousand-odd coins; it was never a big deal. But expectations got wrecked and MicroStrategy became a hunted target: you’re going to sell next month, so I have to sell ahead of you — huge ETF outflows, everybody front-running. This is a fire scene; it’s about who gets out first.

3. You want his interest; he wants your principal

Liu Feng: There’s another big difference this cycle: the financing instruments themselves. In past bears he used near-zero-coupon converts to buy coins slowly; this round he’s launched a whole string of new things starting with STRC. Some say Saylor is a great corporate-finance innovator; others say he’s the creator of Ponzi instruments. How much ability to raise cash does he have left?

Raymond: Let’s trace it. Earliest was issuing stock, easiest to understand: I’m a pool, you buy my new shares, betting that bitcoin per share keeps rising and that bitcoin will one day reach a million, so you should pay a premium — the simple version of mNAV, a very bull-market story. Compare last cycle’s GBTC: a 30–40% premium at the high, a thirty-odd percent discount at the low, so the range was inherently wide; MicroStrategy’s mNAV peaked around 3x in late 2024 and early 2025, which is extraordinary. In last cycle’s bear he mainly sold converts: borrow money, give you the upside as a call option, in exchange for a negligible 1% or 2% coupon. This cycle he’s become very creative: STRK, STRF, STRC, STRD — K, F, C plus a D, a KFC combo. He also said recently that STRK came out of a conversation with GPT, that AI helped him design it. Maybe a joke, but it shows he spares no effort on financial innovation. Let me flag two hidden points. First, around 2000 MicroStrategy triggered a restatement of financials for excessively aggressive revenue recognition and was forced to redo its statements — he was already good at this twenty-odd years ago; it’s character. Second, the scale is already extraordinary: bitcoin holdings around $50 billion, preferred issued to $15 billion, STRC alone around $10.5 billion, with over $100 million of interest a month. Track record, coin-price reflexivity, interest expense, financial instruments layered on top of each other — the company may not actually be in trouble, but the risk people perceive keeps rising. Even if a company has done nothing wrong, if people feel uneasy looking at it they won’t dare lend to it; and the less they dare lend, the worse the financials get, the snowball stops rolling, and whatever you do is wrong. At times like this it’s best to sit still and do nothing — the more you do, the more you get wrong.

Liu Feng: I’d suggest listeners put their attention on STRC alone: the largest by size, everything else is a rounding error. Raymond, introduce the product?

Raymond: STRC is nominally preferred stock; you can understand it as having bought a bond. STRK and STRF have fixed coupons of 8% and 10%; STRC’s coupon is variable, currently annualizing at 11.5%, which is quite high. The mechanism: if STRC’s price falls, the company raises the coupon to attract buyers and pull the price back to 100; back at 100, it issues more. The cost is having to pay interest continuously, and holders have no redemption right, and the company has never done a buyback to this day. Hence these past two days: from 100 down to a low of 82.50, now back at 88.60, still a long way from the theoretical 100. The second change is in the buyers. MicroStrategy previously leaned toward institutional financing — converts are a hedge fund product, and the buyers were Wall Street’s sharks; the STR series became a retail-leaning product, with many people coming in for the 10% or 12% coupon, packaged as something like fixed income, like a money market fund. Wall Street argues about this a lot: this is the twin sibling of those financially engineered wealth products from the financial crisis. Retail dominating is deceptive; they don’t necessarily make prudent decisions. And now it’s de-pegged — you wanted his interest, and Saylor took your principal.

4. STRC isn’t a Ponzi, but a three-ball juggle is much harder than two

Liu Feng: STRC has a par of 100, the coupon is set by the company, and it trades on the secondary market, so at the current price the implied yield is close to 13%. It looks like a bond, but there’s no repayment obligation. Mr Zheng, is this great financial innovation?

Zheng Di: First to add: the STR series has five, F, C, K, D, plus a euro-denominated E which is very small, and their liquidation seniorities all differ. Look at the history: two cycles ago MicroStrategy levered up by issuing ordinary bonds with coupons in the sixes, as high as eight, which he disliked; last cycle he turned to converts — Saylor explained that bitcoin has no carry, so paying interest is troublesome. Converts pay basically 0 or 0.25, and the core is hoping you convert; if you can’t convert he has to repay and roll it. The current $8.2 billion of converts is last cycle’s legacy, $6.7 billion after buying back $1.5 billion. This cycle he’s essentially abandoned zero-coupon converts — I don’t know whether rolling the issuance got hard, or he disliked diluting shareholders, or he was tired of the delta-neutral arbitrage institutions pressing the share price — and shifted at scale to perpetual preferred. Earliest was STRK, authorized at $21 billion in tribute to 21 million coins, with high hopes on it; STRK is low seniority but convertible, a natural extension of the converts. Then he wasn’t satisfied and issued the higher-seniority STRF, then STRC. The full ordering: converts highest, then STRF, STRC, then STRK and STRE, then STRD, and finally common. He’s using a mezzanine-like high-seniority instrument to amplify leverage. The market hasn’t actually analyzed this seniority precisely: the claims ranking ahead of STRC are only about $8 billion, plus STRC’s own $10.4 billion, under $19 billion in total, against a bitcoin reserve of forty or fifty billion — statically it’s fairly safe.

Liu Feng: That’s important information; a lot of analysis is off the top of the head and has never ranked these debt-like instruments.

Zheng Di: But Raymond is right that this is the chained-ships stratagem: stock, coin price and STRC bound together, and one fire may burn the whole chained fleet, forming a downward spiral. So the smart move is not to start at all — which is why the error is the $1.5 billion convert redemption: you think you happily made a hundred million dollars, and in fact you lit the fire. Statically $19 billion against forty or fifty billion is very safe, but once the market panics the collateral itself is shrinking. What the market ultimately tests is whose interests you put first: if you protect STRC, the market concludes you protect neither the share price nor the coin price. So STRC is the fuse for the whole system: stabilize it and everything else has a chance; fail to and the doubts don’t stop. With a coupon of 11.5% and implied yield of 13%, it means the next raise has to be to 13% before people accept it; and the worse case is raising to 13% and still not getting back to 100 — that’s the market’s risk tolerance falling further, where the higher the coupon the more dangerous you look. His share premium could previously sit above 22% largely because of the bitcoin-per-share neutrality policy announced in May: don’t sell stock to pay interest below 1.2x. Before that, the system everyone assumed was: sell stock to pay interest when there’s a premium, sell STRC to buy coins when STRC is near 100, and if it de-pegs to 98 or 97 just wait for it to return to 100 — that logic doesn’t work now. And frankly, do MSTR shareholders genuinely care about bitcoin per share? If I held MSTR I wouldn’t; what I’d care about is bitcoin going up.

Raymond: I don’t care either.

Liu Feng: But that’s the core story for selling stock.

Zheng Di: It is the core story, and people who genuinely treat bitcoin as faith are moved by it; I concede that. But having got to today, it shouldn’t be the primary pitch anymore — it terrified the market: below 1.22x you won’t sell stock to pay interest, so you can only sell coins; and once you start selling, mNAV is permanently below the threshold, the market concludes you’ll keep selling, so it keeps falling, until you prove by action that you don’t apply it rigidly. Several things stacked together. But I don’t think this is STRC’s problem — STRC as financial innovation is genuinely rather good. Why did he arrive at an adjustable coupon? He went to pitch institutions, to pitch pension funds: you don’t dare buy my stock, but surely you can consider a perpetual at 10%? It isn’t directly collateralized by bitcoin, it’s implicitly collateralized, and bitcoin is a quality asset — which is unlike the CDOs of 2008. Financial engineering is fundamentally neutral, like a knife that can save or kill; what actually caused the financial crisis was putting toxic underlying assets in, like subprime. Slice a CDO out of high-rated corporate bonds and there isn’t much of a problem. The core is always the underlying asset — as long as you accept bitcoin is a quality asset (our listeners may not all agree, but everyone here today, and Saylor, and the institutions he pitches, do), a perpetual implicitly collateralized by it has vitality. The pension funds’ rebuttal is also reasonable: a 10% coupon is great, but institutions can’t take large price swings; if you can pin it at 100 we’ll consider it. And so STRC was born: adjust the coupon monthly to keep the price hugging 100. On holder composition, Saylor himself has said STRC is about eighty percent retail — eight billion or so in retail hands, institutions taking twenty-odd billion; whether there are pension funds I don’t know, but I’d guess so. In a sense he’s already achieved an enormous success: implicitly packaging bitcoin as a credit asset and selling it in a perpetual wrapper. Let me also distinguish fact from narrative — long term is decided by fact, short term by narrative. Last year bitcoin broke below its six-year uptrend line against gold, and that was a narrative kill: people who understand know it was a technically indefensible pseudo-narrative, but explaining it to the public is like Kong Yiji saying there are four ways to write the character 回 — nobody follows. The version the public follows is “elliptic curve cryptography can be broken on-chain, bitcoin is dead.” In March this year the narrative changed to “Middle Eastern princes are buying bitcoin frantically,” and after the war started bitcoin was markedly stronger than the Nasdaq and gold and reclaimed the trend line. Is that what happened? I genuinely researched this — it has nothing to do with the princes; it was STRC’s volume explosion: ten billion dollars bought in from March to now. MicroStrategy held bitcoin up. So I genuinely don’t think STRC is a Ponzi; it’s financial innovation. But it made the game harder: previously you only had to keep two balls in the air, the share price and the coin price; now it’s three, and the technique has to be more superb — no wild experiments, no speculative buyback for a ninety-odd-million paper gain. I’ve always liked Saylor and don’t understand why so many people call him a Ponzi. But today I have to say: from the $1.5 billion convert redemption to the bitcoin-per-share neutrality experiment, he made errors.

5. Even HSBC’s perpetuals fell to 30; nobody is coming to save you

Raymond: A follow-up. Is your advice to Saylor to hide in a cave for twelve months and do nothing until the market comes back? Or to keep pushing aggressively — would you, for instance, advise him to sell stock to buy back STRC? By the bitcoin-per-share arithmetic that should also be accretive.

Zheng Di: He needs to be cautious now and can’t blindly keep accumulating BTC.

Liu Feng: Raymond’s questions all carry an implicit assumption about how BTC’s price moves. Before this decline, most of the market thought this cycle wouldn’t have the old kind of drawdown, and the maximum drawdown is now 50%. I lean toward Raymond myself: limited downside, not for too long. Listeners should understand that Mr Zheng’s answer is based on that assumption.

Zheng Di: But that’s only our assumption; running a company can’t be a gamble, and you have to make extreme assumptions: what if the AI cycle makes the US economy great again, and before the mass layoffs arrive that period stretches from six months or a year to two years? That’s exactly what we see now — AI infrastructure pushing GDP and the business cycle up, liquidity compressed, layoffs not yet here. In Saylor’s position I’d conclude the crux is STRC and the cash-reserve panic, and go rebuild the cash reserve: push it back above 12 months as fast as possible, still five or six hundred million dollars of liquidity short. Last week its mNAV premium slid from around 20% to 13%, and I have reason to suspect it’s already selling stock to replenish reserves. If Monday night’s 8-K shows the cash reserve increasing, I won’t be surprised; if it genuinely sold one or two hundred million of stock with all the cash going to reserves, then in my mind it’s already on the path to repair and this episode won’t be severe.

Liu Feng: To confirm: you want to see it raise cash but not buy bitcoin. Is buying coins the wrong choice?

Zheng Di: It already erred slightly last week: it did a $200 million ATM on the stock, spent $100 million on 1,800 coins, and only $100 million went to reserves — that action is what triggered STRC’s further decline.

Raymond: Is it possible he wanted to signal to the market that bitcoin is undervalued, and the market simply didn’t buy in?

Zheng Di: I agree with his intent, but you only needed to buy back those 32, or buy slightly more, say 100. Having money, and even breaking your own neutrality policy to run an ATM, and then not putting the proceeds into reserves but into coins — if you’re an STRC holder you’re certainly panicking. STRC was very likely pushed down by panic sellers in very thin volume, and the company didn’t buy back either. In tomorrow’s 8-K I hope to see it sold another $200 million of stock with all of it going to cash reserves.

Liu Feng: Our recording timing isn’t great — it’s Sunday afternoon, June 21, and what’s genuinely worth watching is tomorrow’s pre-market 8-K. Mr Zheng’s judgement: the good choice is selling stock to replenish reserves; some in the market also suggest selling bitcoin, which Mr Zheng thinks unlikely; and of course he may do nothing at all.

Zheng Di: Selling coins again would be a fresh round of panic; this floor can’t be tested lightly, and the signal would be too confused — you were buying before and now you’re selling. And one more important thing: the share premium is only 13%, so the room to sell stock into reserves is already limited — I remember it hit lows of 5% or 6% — so seize the moment and sell quickly. If tomorrow night’s 8-K doesn’t show stock sold into reserves, I’ll be fairly disappointed: otherwise why did the premium fall to 13%? Either you’re selling, or the market is front-running.

Liu Feng: Back to STRC’s discount. On the roadshow Saylor kept stressing this is a product pinned near par by adjusting the coupon, and it fell into the 80s — his intent wasn’t realized. How do you read that?

Raymond: My reading differs slightly from Mr Zheng’s: this actually shows Saylor’s success is incomplete. First, what he genuinely wanted was the big money in institutional fixed-income pockets — pensions, sovereign funds — and seventy or eighty percent of the book is retail. Below expectation. Second, if I were a pension fund manager and saw a preferred claiming to be pinned at 100 fall to 83 over a weekend, I wouldn’t touch the STR series again for a long time — I wanted high yield and my principal is taking losses. Third, he sold to precisely the wrong people: people who genuinely think bitcoin is good just buy bitcoin; and people who don’t believe in MicroStrategy don’t believe in bitcoin either, in whose eyes it’s a toxic asset. MicroStrategy over the past year hasn’t expanded the circle for bitcoin or for the STR series; no new money came in, it’s still churning inside the old pool. So the question is: last night at 83 or 85, why didn’t retail step in? My guess is another fifty or a hundred million would buy it back to 100 — it didn’t happen over the weekend, and whether it gets back tomorrow depends on the bulls’ defensive strength. Separately, if tomorrow’s 8-K shows Saylor bought coins, I also wouldn’t be especially surprised — the coin has fallen hard and he may feel he’s buying the dip.

Zheng Di: Saylor never buys the dip; he buys coins with something close to physiological inertia: buys every week, buys whenever he has money. That inertia is exactly the source of people’s confidence in him.

Raymond: I strongly agree that stopping buying is better for the whole flywheel — pull back, find a cave and hide for a year. But this man’s character, and the signal he’s transmitted all along, is to keep buying regardless of high or low, easy or hard. If he stopped buying I’d actually be somewhat surprised. Though buying 32, 320 or 3,200 does make a difference.

Liu Feng: So STRC at 88 — is it worth buying the dip?

Raymond: Two perspectives. First, as a traffic light: over the past year STRC has rarely spent three consecutive days below 97, mostly spiking down and recovering immediately; now it’s been under 98 for nearly a week. There’s background here: over a month ago MicroStrategy changed interest payment from monthly to twice a month — because on the 15th, the interest confirmation day, people would bid the price up and then sell after confirmation to arbitrage that dollar, and semi-monthly payment was meant to smooth the volatility. And now it’s in the 80s, which is a red light, effectively an X-ray of the company: the second-most-senior preferred is de-pegging, and people don’t believe this price is stable. Second, whether it’s worth investing in (not investment advice): its seniority is in fact very high, second priority in a company with over $50 billion of bitcoin, and after repaying the converts and STRF there’s still over $10 billion available to it, so bitcoin would have to fall to around $25,000 before it’s touched — an extremely extreme assumption. So on liquidation seniority it’s relatively a good asset; the risk is that short-term panic could de-peg it further.

Zheng Di: There’s another layer: if STRC doesn’t return to 100, MicroStrategy basically can’t buy coins. The structure since March has been: if the share premium isn’t high (read as below 1.5x) don’t sell stock to buy coins, and rely entirely on STRC’s ATM to buy coins; selling stock is for paying interest and building the preferred reserve pool. Selling stock to buy 1,800 coins last week was an abnormal situation, him being anxious to prove he’s a net buyer. This structure can’t be lightly changed: tomorrow it very likely won’t accumulate, and if it again uses stock proceeds to buy coins that only backfires. It needs to stop for a stretch — raise the coupon to 13%, wait for STRC to return to par, and only then continue selling STRC to swallow coins; the rational choice is rebuilding the cash reserve. With the coin price sitting at a relative low, that also implicitly prices in “MicroStrategy can hardly buy in the short term.” One more detail: an investment bank said ETF flows can now explain 30–40% of bitcoin’s daily volatility. On the Friday before the Monday MicroStrategy disclosed selling 32 coins, there was a $1.2 billion sale of IBIT; the CME futures premium to IBIT was very low then, so you can basically judge it wasn’t a basis-trade unwind but an outright sale. I have reason to suspect that $1.2 billion was informed — that they knew in advance MicroStrategy was selling.

Liu Feng: A thought experiment. In the 2008 financial crisis, both Goldman Sachs and Morgan Stanley privately went to Buffett to issue large preferred. Today MicroStrategy has a ready-made preferred book and issuance machinery; if one day it genuinely de-pegged and ran short of cash, is there anyone — certainly not Buffett — who could buy billions of preferred and transfuse it?

Zheng Di: In hindsight, did Goldman and Morgan Stanley actually need that money? No. Buffett’s terms were punishing — a heavily discounted conversion price, a high dividend, redemption clauses, a guaranteed win. So why were they still willing? Because only Buffett could give the market confidence; nobody else would do. Another example: Joseph Lau in Hong Kong in 2008. Do you know how far HSBC’s perpetuals fell? To 30. HSBC is carved into the genes of Hong Kong’s older retail investors; spare money goes into it. Its perpetuals falling to 30 means “it’s about to fail,” and the reason it didn’t fall further is that there were no sellers left, no quotes on the street, and trades happened by telephone. Lau bought HSBC perpetuals at 30 — I don’t know how much — and a year later they were at 90, three times his money. So think about it: for market forces alone to push it back in a tsunami, how far do you have to fall? Even HSBC fell to 30. Goldman’s and Morgan Stanley’s capital in absolute terms was certainly sufficient, but once you enter the spiral and people stop believing you, what you’re short of isn’t money, it’s confidence. If MicroStrategy’s situation genuinely reached 2008-tsunami conditions, I don’t think anyone could step out and resolve it; in the end you’d only have market forces.

Raymond: But is there a possibility that a large financial institution’s balance sheet is fundamentally unreadable, while MicroStrategy’s statements are extremely simple and transparent?

Zheng Di: A big bank’s statements are first hard to read, requiring a lot of footnotes; and second, having read them you still aren’t reassured — how bad are the bad loans really? And they do business with clients, so a crisis deteriorates the revenue side. MicroStrategy’s problem is a different kind: the chained ships drag the coin price down with them. With the sword of Damocles overhead, nobody buys the dip — common in crypto, where a large sell order sitting there unfilled keeps dip buyers away. Saylor’s financial experiment this round effectively tested whether market forces would hold him up, and it validated my view: nobody will. A sale of a mere 32 coins, because it created an expectation of continuous coin selling, and the market refused to support it, forcing him to pivot and correct himself before it stabilized — the halt came exactly at the soft pivot, the moment he announced buying 1,800 coins, when the market discovered he doesn’t apply the neutrality principle rigidly, and the pressure moved from the coin to the stock and the premium fell. If it genuinely reached 2008-tsunami conditions where everyone believed you had to sell tens of billions of coins to repay debt, how could anyone save you? The coin price would certainly collapse and everyone would race to get out first. If I were trapped, I’d run ahead of Saylor and pick it back up cheap after the collapse; still a profit. Trading requires a weak man’s mindset, not a strong man’s: we have no crystal ball, and however well we analyze today, we still don’t know what happens tomorrow. Survival first, run first and look later. If you genuinely want my advice for Saylor: don’t wait for the fire to grow before putting it out. Right now it’s only a rat in the warehouse having knocked over the oil jar and lit a short length of rope; this is the moment to put it out with ten times the effort — get the cash reserve up quickly and the matter disappears at its most primitive stage. Once the fire spreads, it’s hard.

Liu Feng: Both of you think we’re nowhere near the “brink of collapse” that Twitter talks about. But let me ask a retail sucker’s question: under what conditions would the worst outcome appear? What factors would turn MicroStrategy into a genuine black swan?

Raymond: MicroStrategy isn’t an FTX or a Three Arrows that blows up in two or three days; it’s spiral-shaped, iterating in weekly units: get something slightly wrong this week and become 0.9, get something slightly wrong next week and 0.9 times 0.9 becomes 0.81. What actually drives the negative spiral is macro. One point where I differ from the general market view: I think the AI industry bubble isn’t large, but macro risk is very large and far underestimated. Since Kevin Warsh took over, he’s already expressed the most hawkish attitude a new employee can express on day one, and this only leads to an even more hawkish Fed — a new paradigm we all have to adapt to over the coming years. The rate meetings in the second half and the brewing hike expectations will keep pressing risk assets, bitcoin is the particularly weak one among them, and MicroStrategy is under more pressure. My advice is still the same line: hide, find a cave and hide for a year. Borrowing Mr Zheng’s metaphor — I think that in a bear market Saylor is precisely the rat that knocked over the oil jar; he himself is this company’s biggest risk. If Saylor went on holiday for a year the company would be better off. In a bull market, welcome back to work.

Zheng Di: I fairly agree, though I don’t think anything happens to Saylor personally; but right now you do have to be careful: fight the fire with everything, absolutely don’t pour oil on it, this is not the time to keep levering up to buy coins, and he should restore the state before that $1.5 billion buyback. I can understand his anxiety: his typical convert structure is a five-year with investor put rights in year three, so from 2027 the put windows open one after another, and if they can’t convert within three years he may be forced to redeem — from next year there’s rolling refinancing pressure, and that buyback may have been made with this in mind; perhaps the private roll talks weren’t going well, hence this course. But the more that’s true, the more he should put cash flow and debt rolling first, rather than bitcoin per share and continued hoarding.

Liu Feng: Cash is king; that’s his best choice right now.

6. Satoshi would probably send a confused-face meme

Liu Feng: Let’s close on an excellent question Raymond raised earlier: many people today see Saylor as bitcoin’s savior — if Satoshi were still around and saw today’s MicroStrategy, what would his reaction be?

Zheng Di: Satoshi might not necessarily like it, with a reaction roughly like Vitalik’s. Bitcoin becoming on-chain gold was never his intent; what he wanted was peer-to-peer cash. He might feel: you’ve taken an asset with no liabilities and packaged it into a financial company with liabilities and rigid obligations, making it easier for TradFi to buy, but also reintroducing the very things bitcoin was meant to route around — credit, leverage, intermediaries, maturity mismatch. You’ve made bitcoin into the very thing it hated most. Though many things end up quite different from their initial state. Like OpenAI: Musk went to Sam back then because he was anxious about Google developing AI without limit and wanted a non-profit foundation to constrain others; and in Musk’s eyes Sam went from hero to dragon, ten thousand miles from the original intent. From the historical record Satoshi was a very pure, cynical person, distrustful of government and of leverage, and he certainly wouldn’t like this structure. But he might end with one conclusion: whatever the price, bitcoin is still bitcoin, still producing a block every ten minutes. If a big problem does occur, the thing with the problem isn’t bitcoin, it’s you people who built another layer of financial machinery on top of bitcoin — it’s your new Frankenstein that has the problem.

Raymond: If I had Satoshi on WeChat, he’d send me a confused-face meme. Bitcoin was born of the 2008 financial crisis, and from its aesthetics to its technology to its original product design, all of it was to oppose and resist that entire apparatus; and MicroStrategy has hauled the whole apparatus back and used bitcoin as collateral. Deeply ironic. But we’re inside the game and can’t escape either; all we can do is spend two hours discussing when it blows up and whether it’s worth buying. Satoshi designed a currency that required nobody to trust anybody, and Saylor turned it into a machine that requires you to believe in Michael Saylor without limit.

Liu Feng: My first reaction is also that Satoshi certainly wouldn’t like it: bitcoin wasn’t invented to become this kind of complex financial product, resold to different people through Wall Street institutions like snake oil. Today when people talk about how important bitcoin is, they always talk about how high the price is and how big the market cap — none of which is why Satoshi invented it, and none of which is why people who genuinely like bitcoin like it. Greed and appreciation can be part of it; they shouldn’t be all of it. Financial products can be part of bitcoin’s ecosystem; they shouldn’t be all of it. And yet today they seem to have become the most important factor in how people judge bitcoin’s network, ecosystem and price — which is why we spent two hours discussing this, and Satoshi would be somewhat disappointed. Though I also think that if Satoshi genuinely has great wisdom he might just smile and move on: does any of this matter? Even if MicroStrategy collapses and smashes bitcoin’s price to zero, the blocks keep coming, and enthusiasts will keep growing it. Guessing at this is pointless, but I hope the question brings everyone back to the most fundamental thinking: what is bitcoin for, and why do we like bitcoin?

The full 77-minute conversation is available by searching for 苔藓之火 or Web3 101 on Xiaoyuzhou or YouTube.

If you're working on this too — or you think we've got it wrong — write to us at [email protected]; if you'd rather not write, just leave your email below.